A series of experiments were conducted to examine conceptual priming within and across modalities with pictures and environmental sounds. In Experiment 1, we developed a new multimodal stimulus set consisting of two picture and sound exemplars that represented 80 object items. In Experiments 2 and 3, we investigated whether categorization of the stimulus items would be facilitated by picture and environmental sound primes that were derived from different exemplars of the target items. The results demonstrated that the categorization of environmental sounds and pictures were facilitated in a similar way by conceptually related exemplars presented in advance, but only when a long inter-stimulus interval (1000 ms) was used. Additionally, conceptual cross-modal priming effects by picture and sound primes were asymmetric with systematic switch costs across modalities and with differences in the time-course of activation.